SSL_CTX isn't created in accept(), but when a socket on which accept()
is called is created, so it shouldn't be freed when accept() fails, as
this makes the socket unusable (any subsequent operations using SSL_CTX
would be using freed memory).
This caused the segfaults reported in issue #2681, where the second
connection would crash the application.
Signed-off-by: Konrad Sztyber <konrad.sztyber@intel.com>
Change-Id: I1a01a748c5a34ce3dd0fd3c557b860c0ff314b85
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/14355
Reviewed-by: Aleksey Marchuk <alexeymar@nvidia.com>
Reviewed-by: Shuhei Matsumoto <smatsumoto@nvidia.com>
Reviewed-by: Dong Yi <dongx.yi@intel.com>
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: Mellanox Build Bot
Bumped up the priority of the posix and uring sock implementations to
make sure they are selected before SSL, when no impl is explicitly
specified.
Fixes#2681
Signed-off-by: Konrad Sztyber <konrad.sztyber@intel.com>
Change-Id: Ic8e1e2e13f7bce7ccd746f66087e348677df28d9
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/14354
Community-CI: Mellanox Build Bot
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Shuhei Matsumoto <smatsumoto@nvidia.com>
Note, this change only sets defaults for the ID/KEY,
more specific use cases like NVMe/TCP may set the ID and KEY on a per connection basis.
Also simplify PSK identity string, that isn't NVMe focused.
NVMe libraries using this will need to construct more complicated
identity strings and pass them to the sock layer.
Example:
rpc.py sock_impl_set_options -i ssl --psk-key 4321DEADBEEF1234
rpc.py sock_impl_set_options -i ssl --psk-identity psk.spdk.io
./build/examples/perf --psk-key 4321DEADBEEF1234 --psk-identity psk.spdk.io
./build/examples/hello_sock --psk-key 4321DEADBEEF1234 --psk-identity psk.spdk.io
Change-Id: I1cb5b0b706bdeafbccbc71f8320bc8e2961cbb55
Signed-off-by: Boris Glimcher <Boris.Glimcher@emc.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/13759
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Paul Luse <paul.e.luse@intel.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: Mellanox Build Bot
Some of the options in sock_impl_opts could be different for different
sockets (even if they're using the same impl). However, outside of a
few selected options (recv_buf_size, send_buf_size), there was no
interface to change them.
This change will allow users to change impl_opts on a per-socket basis
when creating a socket. Sockets created through accept() inherit
impl_opts from the listening socket.
Signed-off-by: Konrad Sztyber <konrad.sztyber@intel.com>
Change-Id: I7628ae19def25cef6ffa62aa54bd34e446632579
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/13661
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Community-CI: Mellanox Build Bot
It'll make it possible to change some of the impl_opts options on a
per-socket basis, as well as make it easier to use fields common to all
implementations in the generic layer.
Signed-off-by: Konrad Sztyber <konrad.sztyber@intel.com>
Change-Id: Id3b5e0a0b302fdecc2387d07fb87b75b487dc5c5
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/13659
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Community-CI: Mellanox Build Bot
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Both get_opts and set_opts use very similar macros to achieve almost the
same thing, so it makes sense to extract it to a separate function.
Additionally, it'll be also useful in subsequent patches introducing
per-sock impl_opts, as there will be more places when we want to copy
impl_opts.
Signed-off-by: Konrad Sztyber <konrad.sztyber@intel.com>
Change-Id: I8ab27298d62ea0118463ee945c708acd91aa5104
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/13658
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Community-CI: Mellanox Build Bot
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@nvidia.com>
Instead, we pass MSG_DONTWAIT flag to all recvmsg()/sendmsg() calls.
That way, the IOs remain non-blocking, but the socket is marked as
blocking allowing us to pass flags like MSG_WAITALL, which only make
sense if a socket is in blocking mode.
Signed-off-by: Konrad Sztyber <konrad.sztyber@intel.com>
Change-Id: Ic52162e84aa14efaf6ad0fb3343d289822758e81
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/12592
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Community-CI: Mellanox Build Bot
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Shuhei Matsumoto <smatsumoto@nvidia.com>
It'll make it possible to reuse this code for asynchronous read
requests.
Signed-off-by: Konrad Sztyber <konrad.sztyber@intel.com>
Change-Id: I56dab62587884e2e37fad11b5f0d12df92e175ea
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/12590
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Community-CI: Mellanox Build Bot
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Shuhei Matsumoto <smatsumoto@nvidia.com>
This function processes and completes asynchronous write requests, so it
makes sense to rename it. This is done in preparation for handling
asychronous read requests.
Signed-off-by: Konrad Sztyber <konrad.sztyber@intel.com>
Change-Id: I3f36631dc24a3170204aaaba56f4968be0672fe5
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/12172
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Community-CI: Mellanox Build Bot
Reviewed-by: Shuhei Matsumoto <smatsumoto@nvidia.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
This task is used to receive the socket's error queue to check for
completed zero-copy write requests. The rename is done in preparation
for adding a read task that will actually receive the data from the
socket.
Signed-off-by: Konrad Sztyber <konrad.sztyber@intel.com>
Change-Id: I61bf45e210bb09bd89f3161a75478b41bd8eb070
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/12171
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Community-CI: Mellanox Build Bot
Reviewed-by: Shuhei Matsumoto <smatsumoto@nvidia.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
This patch defines a new function, spdk_sock_readv_async(), which allows
the user to send a readv request and receive a callback once the
supplied buffer is filled with data from the socket. It works simiarly
to asynchronous writes, but there can only be a single outstanding read
request at a time.
For now, the interface isn't implemented and any calls will return
-ENOTSUP. Subsequent patches will add support for it in the uring
module and as well as emulation in the posix module.
Signed-off-by: Konrad Sztyber <konrad.sztyber@intel.com>
Change-Id: I924e2cdade49ffa18be6390109dc7e65c2728087
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/12170
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Community-CI: Mellanox Build Bot
Reviewed-by: Shuhei Matsumoto <smatsumoto@nvidia.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Adding more unit tests using standard openssl
The unfortunate small sleep is needed due to issue:
https://www.mail-archive.com/openssl-users@openssl.org/msg02937.html
Change-Id: I6f55453f12371bec6a402ba4c1d20e21aed73cf4
Signed-off-by: Boris Glimcher <Boris.Glimcher@emc.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/12625
Community-CI: Mellanox Build Bot
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Shuhei Matsumoto <smatsumoto@nvidia.com>
Added new `ssl` based socket implementation.
For now we are using hard-coded PSK and only support TLS 1.3
One can use it via sock_set_default_impl RPCs
Nvme/TCP published secure channel specification (TP 8011)
Which is based on TLS 1.3 and PSK.
So this is a primary but not the oly use case.
Before any SSL connection can be established
we need to create SSL context.
The context should be client/server aware.
Similar to regular sockets, to establish connection
server must call SSL_accept and client must call SSL_connect.
For now I'm using PSK and not certificates since
we aim this for NVMe/TCP TP-8011 which supports only PSK.
Adding certificates later on will be very easy.
The complication with SSL state machine during accep and connect
comes with returned SSL_ERROR_WANT_READ or SSL_ERROR_WANT_WRITE.
According to documentation, call have to be repeated in this case.
Using openssl here for TLS from user space.
openssl also has support for kTLS.
Will be part of the next changes.
openssl doesn't have implemetation for iovec
only basic SSL_read and SSL_write.
So adding here SSL_readv and SSL_writev wrappers.
Tested using:
./build/examples/hello_sock -N ssl -H 127.0.0.1 -P 12345
./build/examples/hello_sock -N ssl -H 127.0.0.1 -P 12345 -S
Also tested using:
nvmf_tgt + sock_set_default_impl + perf
Signed-off-by: Boris Glimcher <Boris.Glimcher@emc.com>
Change-Id: Ie730077c5c581b7e112c18f5f9e1b683015e7b4b
Signed-off-by: Boris Glimcher <Boris.Glimcher@emc.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/12327
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Community-CI: Mellanox Build Bot
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Shuhei Matsumoto <smatsumoto@nvidia.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Many open source projects have moved to using SPDX identifiers
to specify license information, reducing the amount of
boilerplate code in every source file. This patch replaces
the bulk of SPDK .c, .cpp and Makefiles with the BSD-3-Clause
identifier.
Almost all of these files share the exact same license text,
and this patch only modifies the files that contain the
most common license text. There can be slight variations
because the third clause contains company names - most say
"Intel Corporation", but there are instances for Nvidia,
Samsung, Eideticom and even "the copyright holder".
Used a bash script to automate replacement of the license text
with SPDX identifier which is checked into scripts/spdx.sh.
Signed-off-by: Jim Harris <james.r.harris@intel.com>
Change-Id: Iaa88ab5e92ea471691dc298cfe41ebfb5d169780
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/12904
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Community-CI: Mellanox Build Bot
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@nvidia.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-by: Dong Yi <dongx.yi@intel.com>
Reviewed-by: Konrad Sztyber <konrad.sztyber@intel.com>
Reviewed-by: Paul Luse <paul.e.luse@intel.com>
Reviewed-by: <qun.wan@intel.com>
Due to the same reason as transport_ack_timeout for
RDMA transport, TCP transport also needs ack timeout.
This timeout in msec will make TCP socket to wait for
ack util closes connection.
Signed-off-by: zhangduan <zhangd28@chinatelecom.cn>
Change-Id: I81c0089ac0d4afe4afdd2f2c7e5bff1790f59199
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/12214
Community-CI: Mellanox Build Bot
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Reviewed-by: Shuhei Matsumoto <smatsumoto@nvidia.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
The process of matching qpair to poll group is split into
two distinct parts that occur on different threads.
See spdk_nvmf_tgt_new_qpair().
This results in a race condition for TCP between spdk_sock_map_lookup()
and spdk_sock_map_insert(), which are called in spdk_nvmf_get_optimal_poll_group()
and spdk_nvmf_poll_group_add() respectively.
Fixes#2113
This patch picks a hint from nvmf_tcp for next poll group,
which is then passed down to spdk_sock_map_lookup().
When matching placement_id exists, but does not have
a poll group assigned - the hint will be used.
Signed-off-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Change-Id: I4abde2bc9c39225c9f5dd7c3654fa2639bb0a27f
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/10271
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Sendmsg_idx is used in _sock_check_zcopy to check whether idx in
MSG_ERRQUEUE and req's idx match. There is no need to update
it if zerocopy is disabled.
Change-Id: I24bb367e0bff006782d9052470857f6e4db90681
Signed-off-by: Richael Zhuang <richael.zhuang@arm.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/12104
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Community-CI: Mellanox Build Bot
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
We try avoiding write a closed socket by checking if the return value
of recv() is zero. However it is not possible to completely avoid writing
a socket which is already closed by the target.
Repeatedly adding/removing listener in the NVMe-oF TCP target caused
SIGPIPE to the NVMe-oF initiator.
Fix the issue by adding MSG_NOSIGINAL to the flag of sendmsg().
Signed-off-by: Shuhei Matsumoto <smatsumoto@nvidia.com>
Change-Id: I273679c91c4b867792e966b1dc2121f6d2188f16
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/12119
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Community-CI: Mellanox Build Bot
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
If the req's cb_fn will close the socket, there is heap-use-after-free
error if continuing to access sock.
Signed-off-by: Richael Zhuang <richael.zhuang@arm.com>
Change-Id: I88c6adb9d25e52d94b08f53e8ccac611c4d29fff
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/11855
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Community-CI: Mellanox Build Bot
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
There is io error when running NVMe over TCP fio test with uring
socket. It's easy to reproduce the bug with the following
configuration:
target 1 core, 16NVMe SSD, 2 initiators each connects to 8 NVMe
namespaces, each runs fio with numjobs=3.
For if in each round, we inset the sock to the head of the
pending_recv list, and then get max_events socks from head of the
list to process, there is possibility that some socks are always
not processed.
Although there was a strategy to cycle the pending_recv list to make
sure we poll things not in the same order. Such as a list: A B C D E F,
if max_events is 3, then this strategy makes the list is rearranged to
D E F A B C. But it will make this strategy not effective if using
TAILQ_INSERT_HEAD(&group->pending_recv, sock...).
Using TAILQ_INSERT_TAIL(&group->pending_recv, sock...) can fix it.
Signed-off-by: Richael Zhuang <richael.zhuang@arm.com>
Change-Id: I8429b8eee29a9f9f820ad291d1b65ce2c2be22ea
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/11154
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Community-CI: Mellanox Build Bot
To allow SO_MINOR updates on LTS for the whole year it is supported,
the major version for all components needs to be increased.
This is to prevent scenario where two versions exists with matching
versions, but conflicting ABI.
Ex. Next SPDK release adds an API call increasing the minor version,
then LTS needs just a subset of those additions.
Increasing major so version after LTS, allows the future releases
to update versions as needed. Yet allowing LTS to increase minor
version separately.
Disabled test for increasing SO version without ABI change, as
that is goal of this patch. This check shall be removed with SPDK 22.05
release.
This patch:
- increases SO_VER by 1 for all components
- resets SO_MINOR to 0 for all components
- removes suppressions for ABI tests
Signed-off-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Change-Id: Id1a5358882dc496faa5b0b5c9a63b326c378c551
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/11361
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
For IPV6, cm->cmsg_level is SOL_IPV6 and cm->cmsg_type is
IPV6_RECVERR. However these combination was not included.
To clarify the fix check if positive conditions are satisfied and
then reverse the result.
Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Change-Id: I675f4337f383d3526fed1b86794697f41113ed4c
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/10428
Community-CI: Mellanox Build Bot
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
Reviewed-by: Konrad Sztyber <konrad.sztyber@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Options modified on sock after connected is also moved to a
function.
Signed-off-by: Liu Qing <winglq@gmail.com>
Change-Id: I4c2a9ae9858c102764959d055bed208b4b0621d9
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/9516
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Community-CI: Mellanox Build Bot
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
This reverts commit 2cd948c4a6.
This commit caused drop in performance tests.
More info in issue https://github.com/spdk/spdk/issues/2158
Signed-off-by: Maciej Wawryk <maciejx.wawryk@intel.com>
Change-Id: Id5d353535323c79e773e33377af388dae47238cb
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/9510
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
When we rotate the socket in the list, we did not check whether the uc pointer
is NULL, then we cause the coredump when using uc pointer.
When we add a new socket (with pollin event) into the pending_recv list,
we add it into the end of the pending_recv list, then it delays the execution
of the newly socket, and it can cause the timeout especially when a socket
is in a connection phase.
So the purpose of this patch is:
1 Revise the rotation logic to handle the two cases, i.e., (1)sock is in the
beginning of the list; (2)sock is in the end of list. The purpose is
to avoid NULL pointer access, and efficently handle the exceptional case.
2 When there is new pollin event of the socket, we should add socket in the beginning
of the list. And this can avoid the new socket handling starvation.
Since max poll event num is 32 from upper layer and if we always put the new socket
in the end of the list, then starvation will occur if there are many socket connection events.
Because if we add the new socket into the end of the pending list, we will always handle the
existing socks first, then the later coming socket(with relatively pollin event) will always be
handled late. Then in the sock connection initialization phase, it will consume a relatively
long time, then the upper layer connection based on this socket will cause timeout,.e.g.,
ctrlr.c: 185:nvmf_ctrlr_keep_alive_poll: *NOTICE*: Disconnecting host nqn.2014-08.org.nvmexpress:
uuid:af56cce7-2008-408c-a8a0-9c710857febf from subsystem nqn.2019-02.io.spdk:cnode0 due to
keep alive timeout.
[2021-08-25 20:13:42.201139] ctrlr.c: 579:_nvmf_ctrlr_add_io_qpair:
*ERROR*: Unknown controller ID 0x1
Fixes#2097
Signed-off-by: Ziye Yang <ziye.yang@intel.com>
Change-Id: I171b83ffd800539e86660c7607538e120fbe1a91
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/9223
Reviewed-by: John Kariuki <John.K.Kariuki@intel.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Community-CI: Mellanox Build Bot
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
When removing a socket from a sock_group, the recv_task should be
cancelled last, because it can be sent out while cancelling other tasks
(if POLLERR is received). Otherwise, we could end up with outstanding
recv requests from a socket removed from a group.
Fixes#2112.
Signed-off-by: Konrad Sztyber <konrad.sztyber@intel.com>
Change-Id: Ic8e24c210541390dd8bdffe8d3bc4e7dd746d4b7
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/9239
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Community-CI: Mellanox Build Bot
Reviewed-by: Ziye Yang <ziye.yang@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
In order to not affect the loopback test.
Also create a sock_common.c file which can be used by posix/uring
implementation. We do not put such code in sock.c. Because sock.c
is the general layer. Other users may include their own user space
sock impelmentations. So put those common code in sock_common.c instead.
Signed-off-by: Ziye Yang <ziye.yang@intel.com>
Change-Id: I983ec2313119539e6eed2d9f11ba1488c0ed6560
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/8769
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Community-CI: Mellanox Build Bot
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
This patch tries to add MSG zero copy feature.
Though io_uring supports buffer registration, it only
support io_uring_prep_write_fixed. It means only one
registered buffer can be used. It does not satisfy our
current usage mode.
According to this situation, we still use the MSG_ZEROCOPY
flags in io_uring_prep_sendmsg.
Furthermore, this new feature is dependent on the kernel
version, The currently verified version is
kernel 5.12 rc3. So it is not enabled in the default manner.
For example, if you want to use it on the target side, you can
use the following rpc to configure:
./scripts/rpc.py sock_impl_set_options -i uring --enable-zerocopy-send-server
Change-Id: Ie7bb828f466362add94891989ddf0950dccd9e80
Signed-off-by: Ziye Yang <ziye.yang@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/957
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Community-CI: Mellanox Build Bot
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
This parameter is still part of API spdk_sock_impl_opts
structure but it is not used. Keep it to support ABI
compatibility since it is located in the middle of the
structure and removing it may break socket opts initialization
or parsing.
Change-Id: Ib641ad7d965d68bc9ebb65dba531408d88cf6fa1
Signed-off-by: Alexey Marchuk <alexeymar@mellanox.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/8914
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Revise the if case to avoid the assert issue.
Change-Id: I095f3d111423e17abaa1f951fe22efb3d2e851b7
Signed-off-by: Ziye Yang <ziye.yang@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/8872
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Community-CI: Mellanox Build Bot
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
This patch is used to improve the performance when
we need to reorder the list.
PS: Bring the similar operations from posix implementation.
Signed-off-by: Ziye Yang <ziye.yang@intel.com>
Change-Id: I7031b35ddb597730ee160690e8ab9caf9b2b64b7
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/8675
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Community-CI: Mellanox Build Bot
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
When entering the if case to order the list, there is bug should be
fixed. The original code does not address this.
The way this happens is when there is a connection left in the socks_with_data list
between polls and there are enough new events detected that it would exceed the
maximal number of events. A connection is left on this list between polls if it isn't
fully drained via reads by the upper layer on each poll loop.
Currently, the maximal socket event num is 32. Then we did not hit this issue
in our normal test cases. But when you use NVMe-oF tcp target to test which is
described in #2105, there are more than 32 active sockets, and it exceeeds
the maximal num of events of polling (32), so we will trigger this issue.
Fixes issue #2015
Change-Id: I9384476fdba8826f5fe55a5d2594e3f4ed3832ba
Signed-off-by: Ziye Yang <ziye.yang@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/8541
Community-CI: Mellanox Build Bot
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
In uring_sock_create(), we loops through all the addresses available.
If something is wrong, we should close(fd) and set fd to -1, and
try the next address. Only, when one fd satisfies all conditions,
we will break the loop with the useful fd.
Signed-off-by: Zhiqiang Liu <liuzhiqiang26@huawei.com>
Change-Id: I22eada5437776fe90a6b57ab42cbad6dc4b0585c
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/8311
Community-CI: Mellanox Build Bot
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Ziye Yang <ziye.yang@intel.com>
In posix_sock_create(), we loops through all the addresses available.
If something is wrong, we should close(fd) and set fd to -1, and
try the next address. Only, when one fd satisfies all conditions,
we will break the loop with the useful fd.
Signed-off-by: Zhiqiang Liu <liuzhiqiang26@huawei.com>
Change-Id: Icbfc10246c92b95cacd6eb058e6e46cf8924fc4c
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/8310
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Ziye Yang <ziye.yang@intel.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: Mellanox Build Bot
After reading the code in detail, I think that we should
not set pipe_has_data= true and socket_has_data at the same time.
Signed-off-by: Ziye Yang <ziye.yang@intel.com>
Change-Id: I8f9f96b16f4f0e0c585877a0dd687a240252a7cf
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/8283
Community-CI: Mellanox Build Bot
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Move from a single flag indicating that the socket is on the
pending_events list to two flags - pipe_has_data and socket_has_data. If
either flag is true, the socket is on the socks_with_data list.
This is necessary to track enough state to avoid doing extra recv()
system calls.
Change-Id: I65e5701dccb0a5bade19f266f164f26706b110d4
Signed-off-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/7595
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: Mellanox Build Bot
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
When zcero copy send is enabled and used by initiator,
it could significantly increase latency in some payloads.
To enable more fine graing configuration of zero copy
send feature, add new parameters enable_zerocopy_send_server
and enable_zerocopy_send_client to spdk_sock_impl_opts to
enable/disable zcopy for specific type of sockets.
Exisiting enable_zerocopy_send parameter affects all types
of sockets.
Signed-off-by: Alexey Marchuk <alexeymar@mellanox.com>
Change-Id: I111c75608f8826980a56e210c076ab8ff16ddbdc
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/7457
Community-CI: Broadcom CI
Community-CI: Mellanox Build Bot
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Reviewed-by: Karol Latecki <karol.latecki@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
On host side the connections are created and then added to thread's
poll group. Those connections could use different NIC queues underneath.
To route all connections of poll group through single queue a unique
placement id is chosen as group_placement_id and each socket of poll
group is marked with group_placment_id using getsockopt(SO_MARK) option.
The driver could use so_mark value of skb to determine the queue to use.
Change-Id: I06bda777fe07a62133b80b2491fa7772150b3b5d
Signed-off-by: Sudheer Mogilappagari <sudheer.mogilappagari@intel.com>
Signed-off-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/6160
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Community-CI: Mellanox Build Bot
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
We do not want to do any further work on adding
the sock to the group if the epoll_ctl (or kevent)
fails.
Signed-off-by: Jim Harris <james.r.harris@intel.com>
Change-Id: I44b6dc86ce5676aa1b8d6c50b86f22758e4e37fa
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/7594
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: Mellanox Build Bot
suppression
When all of the following conditions are met:
- non-blocking socket
- zero copy is enabled
- interrupts are suppressed (i.e. busy polling)
- NIC tx queue is full at the time sendmsg() is called
- epoll_wait sees there is already an EPOLLIN event
then we can get into a situation where data we've sent is queued
up in the kernel network stack, but interrupts have been suppressed
because other traffic is flowing. This makes the kernel miss the
signal to flush the software tx queue. If there wasn't also already
a pending EPOLLIN event, then epoll_wait would have been sufficient
to kick the system out of this state. But when all of this aligns,
it hangs.
We deal with this by detecting the scenario and calling poll(), which
will force the kernel to issue the pending transmits.
Change-Id: Ifb247159b7de16c8fc72a90f0333f5b421c8bd07
Signed-off-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/6750
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Community-CI: Mellanox Build Bot
Busy pollng using recv() is dependent on kernel socket buffer being
empty. Instead poll() function busy polls hw queues with no such dependency.
Signed-off-by: Sudheer Mogilappagari <sudheer.mogilappagari@intel.com>
Change-Id: I1cb101848d51f7778cdf3d4c015d2d03201bdb37
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/7014
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Since the maps are unique to modules, they can store the group_impls
directly.
Signed-off-by: Ben Walker <benjamin.walker@intel.com>
Change-Id: I7f11db558e38e940267fdf6eaacbe515334391c2
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/7222
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
This allows for different policies per module, as well as overlapped
placement_id values.
Change-Id: I0a9c83e68d22733d81f005eb054a4c5f236f88d9
Signed-off-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/7221
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
There was a fix for this that went into the posix layer, but the
underlying problem is the logic in the nvme/tcp transport. Attempt to
fix that instead.
Change-Id: I04dd850bb201641d441c8c1f88c7bb8ba1d09e58
Signed-off-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/6751
Community-CI: Broadcom CI
Community-CI: Mellanox Build Bot
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
